All images available online: right click here and select Open In New Window to view image page simultaneously.
Visual Testimony:
an
epistemology of scientific illustration
“We ought to talk less and draw more.” –Goethe
How does what we see in an illustration become what we know? This paper aims to give an account of the function of scientific illustration in the acquisition and transmission of scientific knowledge, by casting scientific illustration in a new light: as visual testimony.[1]
It is not far-fetched to consider scientific illustration as testimony. ‘Testimony’, as the term is employed in a philosophical sense, does not simply denote a legal witness statement: rather, it concerns the hundreds of beliefs taken on trust from human sources. In this sense, testimony is ubiquitous, shaping our understanding of the world around us. Like verbal testimony, visual testimony offers a view of the world as the informant sees it with the intended result of belief on behalf of the viewer; images similarly shape our understanding, whether teaching, guiding, illustrating, supporting, or re-presenting. However, visual testimony is sufficiently distinct from its verbal counterpart as to deserve special examination. This paper attempts to elucidate some of the ways in which visual testimony deserves particular epistemological attention with respects to the viewer’s management of an informant’s visual claims.
Why re-frame a study of scientific illustration as testimony? The need for a new critical structure for the examination of images in science is deeply felt. Scientific illustrations have been a subject of academic attention at least since Martin Rudwick’s 1976 article on visual languages in geology; since then, historians and philosophers of science have critically addressed visual representations from several angles.[2] Yet so far, the theoretical literature is disparate and unfocussed. Few attempts have been made at an epistemology of scientific illustration beyond the recognition that images should not be mistaken for the actual object represented; this approach locates epistemic interest exclusively in the image’s relation to reality,[3] leading many authors to place an early illustration beside a current photograph of the object as if to evaluate it for accuracy. Such historical insensitivity breeds frustration and does not, I believe, produce a better philosophical or historical understanding of scientific illustration.
Framing scientific illustration as a kind of testimony, however, redirects, focuses and unifies discussion and analysis concerning images in science. The definition of testimony necessarily relies on the testifier and the recipient, allowing a multivalent relational framework incorporating informant-object, informant-image, viewer-image, and viewer-object interactions; each of these relationships allows space for the models already offered in academic literature to be employed side by side in historical case studies or epistemic explorations. The testimonial framework, further, locates epistemological discussion of the image as mediator of reality, without engaging in debates as to what that reality is. Such an approach offers fascinating epistemological questions about imagery and its role in scientific knowledge-production, while distancing the discussion from issues of ‘the real’, finding a common constructivist ground on which historians and philosophers can comfortably work.[4]
Within a testimonial understanding of images, epistemological questions must address what is being asserted, how it is evaluated or managed, and how it is convincingly conveyed, while taking into account historical conventions, viewing relationships, and the scope and development of visual language systems. This would enable an epistemic analysis of an image as testimonial which can ask, in the end, about both what information the illustration seeks to convey and the strategies it takes to ensure acceptance of its testimonial content. A full account of the scientific illustration as testimony would address these issues of content and management equally, but such an account would require several chapters, even in the form of a simple programmatic sketch. The scope of this essay has therefore been narrowed to focus on the management of visual epistemic claims. The present question, then, is this: in what special ways do images convince the viewer of their truth-content? Put differently, what does the viewer require of a scientific illustration in order to favourably receive its testimony and incorporate the informant’s view of the world into her own?
In ascertaining the truth of a verbal claim, the listener brings diverse considerations to bear on the speaker’s authority or the claim’s own accuracy: these have been identified in such guises as reference to previous experience, gentlemanly conduct, or inference to best explanation.[5] The image must also convince the viewer of its truth-content, and the viewer’s expectations, policies and methods of management play directly into the strategies employed by visual testimony in order for the illustration to be favourably received. Yet methods for claiming truth-content within the visual sphere differ from those employed in verbal claims. In particular, the image may efface or highlight the role of the informant, confuse judgements of competence, make assertions that (literally) appear contextually plausible and unproblematic, and provide lovely explanations. This essay will explore the scheme for testimonial management presented by visual in comparison to verbal testimony, highlighting a particular set of cues by which the informant communicates her view of the world to the viewer so as to encourage confidence in the information exchanged.[6]
One of the most significant contrasts between visual and verbal testimony is the visual’s special appeal to observation. Both forms of testimony are, in a sense, observed: we must see or hear the testimony itself. However, images can effectively give the impression that what is perceived is not testimony about p, but p itself. ‘Evidence that p’, visually speaking, is unusually self-effacing: it is made to ‘look like’ empirical instead of testimonial evidence. The artist can easily make herself invisible, presenting the image as a frame through which the viewer can directly observe the real world, and making the viewer a virtual witness of an object or event.[7] The resulting spatial collapse between the viewer and the object sets up a viewing relationship in which there appears to be no informant, and the picture’s status as mere assertion is overlooked. So while a listener may know she is taking someone’s word for p, the viewer, in seeing an image of p, believes he has a privileged perspective on p similar to first-hand observation. Reading about the eruption of Vesuvius is a different experience than seeing paintings by its witnesses (Fig.1): seeing a picture can really seem like ‘being there’.
Unlike verbal testimony, then, which can never truly escape the frame of reportage, the visual evades a traditional testimonial framework, making the viewer feel as if they are making a direct observation, confusing the line between what is known by report and what is known by personal observation[8] in a way that verbal testimony cannot hope to accomplish. This is not limited to use of naturalistic or ‘tromp l’oeuil’ illustrations; even in abstract visual languages, relationships between data sets become concrete when presented in visual form. These are also absorbed as though their information is observable first-hand: we ‘see’ topography in a two-dimensional contour map, or the hydrogen content of a planet in a processed astronomical image. In contrast, verbal description of a topographical plane – in spite of the use or rejection of colourful adjectives or figures of speech – cannot provide as full or detailed an account of the important details of the scene as a map or landscape sketch, and an oral or written form of testimony will always appear to the listener to be tainted by the informant’s bias or point of view.
The empirical appeal of the visual
carries particular implications when a scientific illustration makes observable
what is otherwise unobservable without the use of instrumentation. This may be the quantum world of the
electron, a God’s-eye view of a continent, the space inside the human body, or
a computer enhanced nebula;[9]
it may also be objects that are not physical bodies at all, such as depictions
of solar magnetic fields (Fig.2). The result of viewing such images is
again a spatial collapse, this time of human and instrumental informants. The
viewer again assumes a privileged empirical perspective, although it may be
physically impossible to ever ‘be there’.
Visual testimony can only truly seem like being there if a significant condition is fulfilled: that of the informant’s reliability (I-R). A viewer of a scientific illustration must first be convinced of the illustrator’s reliability as a good informant before she allows herself to be a virtual witness when viewing his illustration. Informant reliability is contingent upon the nature of the informant’s relationship to the represented object; in scientific illustration, this relationship is judged with an eye to the informant’s positioning with respect to the object, and the informant’s own credentials or expertise.
To fulfil the I-R condition, the informant must appear to be sincere to the object: she must encourage the viewer’s trust in herself as original viewer. In this respect, the role of the instrument or the artist as instrument becomes a site for discussion. Not only must the informant claim a proper use of scientific instrumentation, but she must also claim a proper use of herself as instrument to ensure that the represented phenomena will be taken as trustworthy representations of the world.
Correct use of instrumentation is not simply dependent on the ability to employ an instrument in scientific research; rather, it refers to the informant’s conscious position with respect to an instrument in order to have access to truth about objects in the world.[10] This correct positioning is by no means static, but is historically contingent alongside concepts of ‘seeing’ and reliable witnessing within a changing psychology of perception. For example, early modern communities believed all that was needed for accurate image production of microscopic observations was a ‘sincere hand and a faithful eye’; later observers were suspect of the human as accurate instrument and effaced the informant from the process in an appeal to the purity of technical and mechanical representation. The claim to proper positioning may be made textually explicit, as in Hevelius’ or Hooke’s extended passages dedicated to observational methodology, superior instrument-making, and familiarity with the object bred over long periods of careful examination. It may be claimed within the image itself, sometimes with the figure of the artist in the picture busily engaged in drawing the presented landscape. It may even be implicitly claimed in the choice of medium, such as the photographic plates which, for a Victorian audience, gave sole access to perfect and exact reproductions of phenomena.[11] In each of these cases, the appeal is made to a well-judged and proper use of instrumentation in order to access and reveal truth about the object being explored.
It is further crucial that the scientific artist demonstrate a proper understanding of the practice of representing. In the delicate process of seeing and representing, the artist becomes an instrument herself, attempting to capture and relay important information about the object in illustration. It is therefore critical to her testimonial success that she become the right kind of instrument. Approaches to seeing and representing are broad and varied: should the object ‘speak for itself’, must the artist depict a universal type, or is interpretive intervention required to highlight the important aspects of the picture? The correct choice for an historically situated informant will necessarily be historically contingent, as different eras, disciplines, and media make different demands on scientific illustrations. The correct choice, in the end, will make the informant appear a reliable instrument to the viewer, a good lens through which to gain accurate information about the world.
Once the informant has established (implicitly or explicitly) a trustworthy approach to seeing and representing, his perceived role in the image’s production changes. He can thereby step away from the unreliable role of ‘artist’ – subjective, interpretive, or otherwise decidedly non-scientific – to become a reliable scientific informant in the pure and correct transmission of true information about the world. Therefore, this emphasis on the informant’s role does not compromise the establishment of virtual witnessing; rather, it strengthens the perceived ability for empirical access, as the viewer will only trust the images she sees as tantamount to ‘being there’ for artists she trusts.
Informants may also appeal to their special credibility as competent source of scientific information in order to satisfy the I-R condition for their viewers. In the visual as in the verbal sphere, the informant’s credentials indicate her unique competency with the information at hand, whether through immediacy or training, and serve as cue to the viewer judging her testimony: we are more likely to trust a university professor when she speaks about cell reproduction than our next-door neighbour. However, it is difficult to establish scientific competence directly in a visual medium. Instead, the viewer may extrapolate from a display of observational or representational competency to infer the informant’s level of scientific skill. Yet while these displays may sometimes prove good indicator’s of informant credibility, acute artistic observation does not always guarantee acute scientific observation, and representational mastery is an inconsistent indicator of scientific training.
A conflation of scientific and artistic observational skill is understandable: an illustration’s complexity and mastery communicates the artist’s keen eye and sound observational practice, two features believed common to art and science. The meticulous care and attention required to produce Hooke’s Micrographia or Hevelius’ Selenographia (Figs.3&4) is indeed meant to indicate these informants’ perspicacity in scientific observation and expertise with regards to the object at hand. But it is very easy to produce a picture that appears to be the result of careful observation. Naturalistic representational strategies such as perspective, chiaroscuro, or the use of a camera obscura quickly deceive the viewer into believing that an illustration is not just drawn well, but drawn perfectly, from perfect observation.[12] Thus, the techniques employed to produce Galileo’s moon, Dürer’s rhinoceros, Leonardo’s anatomies, or Stubbs’ moose (Figs.5-8) convinced generations of viewers of true-to-life illustration, resulting in a profound impact on subsequent viewings, representations and interpretations, although the actual viewings of the objects they propose are not physically possible.[13]
If demonstrations of observational competence are an unreliable indicator of scientific competency, representational competence proves even less reliable. In scientific illustrations, visual language competency is implicitly identified with scientific skill, whether or not this is actually the case, and the more convincing the illustration, the more convincing the testimonial. This inference is easily observed in naturalistic illustrations, where the artist may draw the object (theoretical or otherwise) so realistically as to make it come alive for the viewer. Yet a naturalistic style is not mandatory: mastery of more abstract visual languages can fool the eye into believing as well. Paradoxically, the more abstract or unfamiliar the visual language and the more data it relays, the less critical the viewer may become of its assertion: consider the data sets smoothed into an image of solar magnetic fields (Fig.2) or the fictitious elephant-shape mountain embedded in a busy contour map (Fig.9), in which both sets of informants hoped to avoid detailed critical analysis of the data presented in visual form.[14] The viewer’s analytical oversight here is not due to an overload of visual information, but rather to his visual assessment of the informant’s competence with a specialised visual language as direct indicator of scientific trustworthiness. The result is a perceived ‘expert opinion’, which the viewer may not feel he has the appropriate skills to judge critically.
When an illustration ‘looks like’ expert opinion, the viewer assumes that it was indeed produced by an expert. However, the scientist is rarely the author of his own illustrations: a fact often overlooked in our daily processing of scientific visual testimony. Judging the role of the artist employed by the scientist in testimonial management is complex, as it is governed by a set of historically contingent ‘correct’ relationships in the production of illustrations: between artist and scientist, expert and non-expert eye, artisan and science practitioner. Hooke complained bitterly about Micrographia’s engraver’s inability to produce an exact likeness to the original observations; Hevelius was praised for his capacity to engrave his own plates in situ; specially trained and qualified ‘medical artists’ were required for anatomical textbook illustrations in the twentieth century.[15] When the influence of representational competence with regards to testimonial management is considered, the changing relationship between scientist and artist takes on particular significance within historical and philosophical inquiry.
Artistic demonstrations of observational or representational skill, therefore, are not always trustworthy indicators of scientific skill or honesty. But the viewer, seeking evidence of her informant’s credibility, continues to conflate these displays of artistry with scientific competence. This confusion between illustrative gloss and scientific substance leads to a comparison with verbal testimony in the role of glibness of speech or rhetoric; in both cases, the informant resorts to form and style, not content, to convince his audience of his testimonial’s veracity.[16] A full discussion of visual rhetoric is not possible within the scope of this paper, but the comparison is interesting within the field of testimonial management; it indicates the complexity of judging informant reliability based on subject competency, and pinpoints the way in which the image contains considerable convincing power in both how it reveals and what it effaces. In any case, when representational skill is employed to support the informant’s claim to credibility and establish a situation of virtual witnessing, the I-R condition may be unjustifiably fulfilled.
The decision of whether or not to accept visual testimony is not entirely based on the informant; much influence is proper to the image itself. This is, of course, true of verbal testimony as well: testimonial statements must normally cohere with previously accepted background knowledge, for example. In addition to its role in knowledge acquisition in empirical situations, Inference to Best Explanation has been proposed as managing device for verbal testimony.[17] As the scientific illustration often contains significant explanatory power, it is not untenable to draw a parallel between inferring to Best Explanations in the verbal and visual spheres.
If verbal Best Explanations must integrate into background knowledge with minimal conflict, the image proves a potent form of testimonial in its ability to place the object concerned directly into a familiar context. As the object is integrated into a scheme of knowledge visually and without discord, coherence does not need to be inferred: it is simply observed. This is particularly clear in the scenes from Deep Time illustrated in the mid-nineteenth century (Fig.10). Although the prehistoric animals may be unfamiliar, their world is not; the genre of landscape painting is exploited in order to situate the creatures on Earth, albeit in wild territory, emphasising not simply their plausibility but even their corporeality.
Scientific illustrations, further, are able to explain or highlight the explanatory content of a scientific theory in a simple, understandable manner,[18] affording another comparison to an inferred Best Explanation, which aims to provide a clear and simple understanding of the object in question. Clarity in general is upheld in the persuasive presentation of scientific information, and visual exposition is often employed as a good method to make an explanation clearer. Mechanical diagrammatic drawings serve as excellent examples: relationships between parts are mapped or laid out in a manner that explains the machine’s overall function (Fig.11). Illustrations in organic chemistry or biology textbooks attempt the same style of explanation, mapping out an exchange of atomic particles or chemical compounds in a chemical reaction process. Such diagrams graphically provide an understanding of the object by mapping out the object’s function in specific relation to its form. This is highly reminiscent of not just a Best Explanation, but what Lipton has called ‘lovely’ explanations.
Lovely explanations are distinguished from likely explanations in the process of inferring to a Best Explanation: while likely explanations have plausible truth content, lovely explanations establish neat causal relationships between otherwise disjointed or competing information. Interestingly, an explanation’s loveliness is often perceived as a guide to judging its likeliness in the pursuit of a Best Explanation. Visual testimony has specific strategies at its disposal to make an explanation look especially lovely and therefore highly likely, thus possessing a particular ability to trigger an Inference to the Best Explanation.
Firstly, a lovely explanation may appear likely because the image makes an explanation natural. This does not refer to the use of naturalistic representation; rather, what was theory or hypothesis is mapped out onto the represented object itself, making the theory concrete. This mapping, in turn, structures subsequent modes of seeing, understanding and representing. Consider the popular model of the atom: atomic illustrations have mapped current hypotheses onto the object to the point where the hypothesis is the object, even if this hypothesis has since been rejected (Fig.12). Similarly, as individual identities are subsumed into the creation of an ideal type,[19] the latter’s delineation is dictated by overarching theories of form and function to such an extent that the universal or perfect specimen as structured by dominant theory is the object. Enlightenment drawings of ideal skeletons mapped contemporary concepts of masculinity, femininity, function, and even classicism onto their objects to the point where those concepts became the skeleton’s defining characteristics (Fig.13).
In these cases, what was formerly an explanation or hypothesis becomes conceptualised and naturalised as the thing itself. This practice is, I would argue, part of a vertical inference process: i.e. inference from what we have observed to what we cannot observe. We may infer the existence of neutrinos from faint traces in underground deuterium tanks, but it is the illustration of these neutrinos, their possible structure or their reaction with deuterium, that makes them concrete for the viewer. From a few scattered dinosaur bones, we draw fantastical creatures that no human being could observe first hand; we similarly draw convincing atomic structures or distant solar systems on the basis of limited evidence. As we do so, we draw a vertically inferred theory into life, thereby strengthening the inferential link between the observed and the unobservable. When we draw a theory about an object into credible existence and, further, when this illustration becomes or defines the object itself, it is not an understatement to say that the illustrated explanation appears extremely likely.
Secondly, the notion of a lovely explanation takes on new meaning within a context of artistic and scientific aesthetics. Classicism in eighteenth century anatomy (Fig.13) provides one instance of aesthetics’ intrusion into scientific art, but the aesthetic carries actual convincing power when image-loveliness is equated with theory-loveliness. Illustrations of the gently coiling spiral of the double helix made it a very lovely (i.e. visually appealing) explanation for the structure of DNA, though this appealing simplicity and grace are not perceivable in the accompanying written theory (Fig.14). Watson’s frank account of the discovery of the double helix further highlights the importance of visuals in the race between Pauling’s laboratory and his own enterprise with Crick (Fig.15); it was crucial for both research teams to present an aesthetically appealing visual demonstration of their theory, whatever the complexity of the content.[20] Again, lovely (in this case, aesthetically lovely) aspects of the scientific explanation were assumed to be a sign of its theoretical likeliness.
Finally, a lovely explanation can become the Only Explanation if it successfully floods the visual medium. The proliferation of an illustration in the public domain makes it the only image of an object which the public might see, and therefore the definitive one: our genetic code looks like and is a double helix, atoms look like and are mini-solar systems. The rise of such illustrations to iconographic status is reliant on both the above-discussed naturalisation of vertical inference and the aesthetic appeal; the image must stand for and be the thing itself, and must be lovely enough to be recognisable, easily reproducible and ‘best-selling’. Such popular science illustrations come to hold iconographic meaning within the public sphere, and this mass circulation makes the image trustworthy not as the likely, but the only scientific explanation for the viewer.[21]
When an image becomes the Only Explanation, it sets the standard that other observations and illustrations must meet. We all know what atoms and DNA ought to look like, but less simplistic pictures also become prescriptive. For example, Hooke’s famous illustrations of the tick, the fly’s eye, and the ink blot rose in status through extensive reproduction; their assumption, appropriation and reproduction in textbooks into the twentieth century came to structure the empirical experience of microscopy, indicating what the objects ‘ought to look like’, and therefore – in order for the learning practitioner to achieve the same results – how microscopy ought to be done (Figs.16&17). Similarly, the first telescopic images of the moon by Galileo set the standard for telescopic observation, as well as for lunar seeing and drawing (Fig. 18).[22] Such examples demonstrate the extent to which imagistic testimony, through mass reproduction and circulation, shapes not only visual culture but also personal experience.
In this discussion of lovely, likely and only explanations, it is important to remember that scientists judging theories as such does not require analysts to make any claim to their actual truth content. The advantage of examining images as visual testimony is precisely the ability to distance analysis from what is actually true. The above-proposed analyses of images of DNA, dinosaurs, or atoms do not assume that such pictures are or are not ‘true’; rather, they aim to explore the viewer’s judgement and subsequent understanding, based on an illustration, of what reality is like. The role of the lovely explanation in Best Explanation choice, then, serves to highlight the scientific illustration’s power to convince viewers of their theory’s likeliness and truth-content. As such, these potent examples demonstrate the extent to which imagistic testimony, for better or for worse, shapes our understanding of the world around us.
In comparisons with verbal testimony the particulars of the visual become clear. While scientific illustrations, like verbal testimony, seek testimonial validity in appeals to reliability, competency, and explanatory quality, they have a particular set of tools at their disposal in order to convince the viewer of their truth-content. Firstly, an appeal to direct observation is made possible by the effacement of the informant and the confusion of an illustration of p with p itself. Secondly, virtual witnessing is contingent upon the informant’s reliability (I-R), a condition which can be satisfied (but not always justifiably so) by an appeal to correct use of instrumentation or by visual appeal to scientific competence. Finally, the image has considerable explanatory power, and is uniquely able to make a lovely explanation appear a likely, Best or Only Explanation. With such a range of capacity, it is no wonder that a picture is often said to be worth a thousand words.
Scientific illustrations not only play a strong role in the acquisition of scientific knowledge, but are also central parts of our material culture and our understanding. This paper has attempted to outline some of the ways in which images function as testimony in the hopes of building an epistemic account of scientific illustration in this broad and influential role. As the footnotes demonstrate, many of the examples and points within this essay have been discussed before. Previous discussion, however, has often taken place within a context of the problematic examination of scientific images’ relationships to reality. It is now common knowledge in science studies that scientific drawing is a complex process, and it is time to incorporate such knowledge into a new mode of viewing scientific illustration.
The structural framework proposed by an approach to images as visual testimony rearranges and reviews previous scholarship in, I believe, an advantageous manner. Moving discussion away from how or if people can draw what is real allows a more thorough epistemic and historical approach to the topic. The appeal to testimony, then, is an attempt to locate the study of images between the world of objects and the world of testimony in which we live, to see how images transmit, mediate or perpetuate ideas about the world around us. Discussion can now centre around visual management: the judgement of what illustrations we accept and why. The analysis of images provided, therefore, should not only elucidate the role of illustration in the transmission of scientific knowledge, but also enable further inquiry into both particular historical examples and into the philosophical role of images as a whole. A recognition and subsequent re-cognition of scientific illustration as visual testimony, I hope, can suggest a new and rewarding critical structure for historical and philosophical inquiry into the role of pictorial representation in science.
Alpers, S. (1983) The Art of Describing: Dutch Art in the Seventeenth Century (Chicago: University of Chicago Press).
Ashworth, W. B. (1990) ‘Natural History and the Emblematic World View’, in D.C. Lindberg and R.S. Westman (eds), Re appraisals of the Scientific Revolution (Cambridge: Cambridge University Press), pp. 303-332.
Bender J., and Marrinan, M. (2002) The Culture of Diagram (Unpublished manuscript).
Bloom, T. F. (1978) ‘Borrowed Perceptions: Harriot’s maps of the moon’, Journal of the History of Astronomy 9, 117-122.
Coady, C. A. J. (1992) Testimony: A Philosophical Study (Oxford: Clarendon Press).
Coady, C. A. J. (2000) ‘Testimony’, in Routledge Encyclopedia of Philosophy. (http://www.rep.routledge.com: accessed February 19, 2002)
Crary, J. (1990) Techniques of the Observer (Boston: M.I.T. Press).
Dancy, J. (1985) An Introduction to Contemporary Epistemology (Oxford: Basil Blackwell Ltd).
Daniels, S. (1999) Joseph Wright (New Jersey: Princeton University Press).
Daston, L., and Galison, P. (1992) ‘The Image of Objectivity’, Representations 40, 82-128.
Dear, P. (1987) ‘Mathematical Science and the Reconstitution of Experience’, Studies in History and Philosophy of Science 18, 133-175.
Dennis, M. (1989) ‘Graphic Understanding: Instruments and Interpretation in Robert Hooke’s Micrographia’, Science in Context 3, 309-364.
Fricker, E. (1995) ‘Telling and Trusting: Reductionism and Anti-Reductionism in the Epistemology of Testimony’, Mind 104, 393-411.
Galilei, G. (1610) Siderius Nuncius (Oakland: Octavo, 2001).
Galison, P. (1998) ‘Judgement Against Objectivity’, in C.A Jones and P. Galison (eds) Picturing Science Producing Art (London: Routledge), pp. 327-359.
Gombrich, E. H. (1960) Art And Illusion: A Study In The Psychology Of Pictorial Representation (Princeton: Bollingen Series).
Grayling, A. C. (1995) Philosophy: a Guide Through The Subject (Oxford: Oxford University Press)
Hacking, I. (1983) Representing and Intervening (Cambridge: Cambridge University Press).
Hacking, I. (2000) The Social Construction of What? (Cambridge, Mass.: Harvard University Press)
Hall, B. S. (1996) ‘The Didactic and the Elegant: Some Thoughts on Scientific and Technological Illustrations in the Middle Ages and Renaissance’, in B.S. Baigrie (ed), Picturing Knowledge: Historical and Philosophical Problems Concerning the Use of Art in Science (London: University of Toronto Press, 1996), pp.3-39.
Hamilton, W. (1776) Campi Phlegraei: observations on the volcanos of the two Sicilies, As They have been communicated to the Royal Society of London (Naples).
Harley, J. B. (1988) ‘Maps, Knowledge and Power’, in D. Cosgrove and S. Daniels (eds), The Iconography of Landscape: Essays on the Symbolic Representation, Design, and Use of Past Environments (Cambridge: Cambridge University Press), pp.277-312.
Hevelius, J. (1647) Selenographia sive Lunae Descriptio (Gedani: Typis Hünefelianis).
Hooke, R. (1665) Micrographia (Oakland: Octavo, 2001).
Hume, D. (1777) An Inquiry Concerning Human Understanding, Harvard Classics vol 37, (Cambridge, Mass.: P.F. Collier & Son, 1910).
Ivins, W. M. (1953) Prints and Visual Communications (London).
Kemp, M. (1990) ‘Taking it on trust: form and meaning in naturalistic representation’, Archives of Natural History, 17:2, 127-188.
Kemp, M. (1996) ‘Temples of the Body and Temples of the Cosmos: Vision and Visualisation in the Vesalian and Copernican Revolutions’, in B.S. Baigrie (ed), Picturing Knowledge: Historical and Philosophical Problems Concerning the Use of Art in Science (London: University of Toronto Press), pp. 40-85.
Kemp, M. (1997) ‘Stubbs’ Seeing’, Nature 390, 348.
Kusukawa, S. (1997) ‘Leonhart Fuchs on the Importance of Pictures’, Journal of the History of Ideas 58, 403-427.
Kusukawa, S. (2000) ‘Illustrating Nature’, in M. Frasca-Spada and N. Jardine (eds), Books and the Sciences in History (Cambridge: Cambridge University Press), pp. 90-113.
Lipton, P. (1991) Inference to the Best Explanation (London: Routledge).
Lipton, P. (1998) ‘The Epistemology of Testimony’, Studies in History and Philosophy of Science 29:1, 1-31.
Lynch, M. (1994) ‘Representation is Overrated: Some Critical Remarks about the Use of the Concept of Representation in Science Studies’, Configurations 2:1, 137-149.
Lynch, M., and Edgerton, S. (1996) ‘Abstract Painting and Astronomical Image Processing’, in A.I. Tauber (ed), The Elusive Synthesis: Aesthetics and Science (Netherlands: Kluwer Academic Publishers), pp. 103-124.
Nuti, L. (1999) ‘Mapping Places: Chorography and Vision in the Renaissance’, in D. Cosgrove (ed), Mappings (London: Reaktion).
Pang, A. S.-K. (1994) ‘Victorian Observing Practices, printing technology, and Representations of the Solar Corona, (1): The 1860s and 1870s’, Journal for the History of Astronomy 25, 249-274.
Pang, A. S.-K. (1995) ‘Victorian Observing Practices, printing technology, and Representations of the Solar Corona, (2): The 1880s and 1890s’, Journal for the History of Astronomy 26, 63-75.
Pang, A. S.-K. (1997), ‘Visual Representation And Post-Constructivist History Of Science’, Historical Studies In The Physical And Biological Sciences 27, 139-71.
Phillips, R. (1811). The Wonders of the Microscope: or, an explanation of the Wisdom of the Creator, in Objects comparatively minute. Adapted to the Understanding of Young Persons (London: James Gillet).
Rudwick, M. J. S. (1976) ‘The Emergence Of A Visual Language For Geological Science 1760-1840’, Historia Scientiarum 14, 149-95.
Rudwick, M. J. S. (1992) Scenes From Deep Time (Chicago: University Of Chicago Press).
Serjeantson, R. W. (1999) ‘Testimony and Proof in Early-Modern England’, Studies in History and Philosophy of Science 30:2, 195-239.
Schaffer, S. (1998) ‘On Astronomical Drawing’, in C.A Jones and P. Galison (eds) Picturing Science Producing Art (London: Routledge), pp. 441-474.
Schiebinger, L. (1989) The Mind Has No Sex? Women in the Origins of Modern Science (Cambridge, Mass.: Harvard University Press).
Shapin, S. (1994) A Social History of Truth (Chicago: University of Chicago Press).
Shapin, S. and Schaffer, S. (1985) Leviathan and the Air Pump: Hobbes, Boyle and the Experimental Life (New Jersey: Princeton University Press).
Tibbets, P. (1990) ‘Representation and the realist-constructivist controversy’, in M. Lynch and S. Woolgar (eds), Representation and Scientific Practice (Cambridge, Mass.: M.I.T. Press).
Topper, D. (1996) ‘Towards an Epistemology of Scientific Illustration’, in B.S. Baigrie (ed), Picturing Knowledge: Historical and Philosophical Problems Concerning the Use of Art in Science (London: University of Toronto Press, 1996), pp. 215-249.
Tucker, J. (1997) ‘Photography as Witness, Detective, and Impostor: Visual Representation in Victorian Science’, in B. Lightman (ed), Victorian Science in Context (Chicago: University of Chicago Press), pp. 378-408.
Tufte, E. R. (1983) The Visual Display of Quantitative Information (Connecticut: Graphics Press).
Watson, J. D. (1968) The Double Helix: A Personal Account of the Discovery of the Structure of DNA (London: Penguin Books).
Whitaker, E. A. (1999) Mapping and Naming the Moon: A History of Lunar Cartography and Nomenclature (Cambridge: Cambridge University Press).
Winkler, M. G., and Van Helden, A. (1993) ‘Johannes Hevelius and the Visual Language of Astronomy’, in J.V. Field and F. A. L. James (eds), Renaissance and Revolution: Humanists, Scholars, Craftsmen, and Natural Philosophers in Early Modern Europe (Cambridge: Cambridge University Press), pp. 97-116.
I am grateful also to Prof. Peter Lipton for his guidance and support, and to Dr. Anjan Chakravartty, Dr. David Chart, Axel Gelfert, Mark Sprevak, and the Biological Sciences Reading Group for their helpful comments and advice.
[1] This paper focuses on two-dimensional visual representations, i.e. images and illustrations. In what follows, ‘informant’ refers to the producer of the image, ‘viewer’ to its consumer.
[2] See for example Daston & Galison (1992) on vision and objectivity; Ashworth (1990), Gombrich (1960), Rudwick (1976) on visual languages and schemata; Alpers (1983), Crary (1990), Pang (1997) on technological influences.
[3] See ‘Break’ section in Hacking (1983), Topper (1996).
[4] I am well aware of the nebulousness of the ‘constructivist’ category indicated by Hacking (2000), but here I offer a response to Lynch’s (1994) call for an ‘explosion’ of the concept of representation and Tibbet’s (1990) hope for ‘a thorough-going constructivist account of representation and [representational devices] in science … which avoids any entanglement with this unresolvable and debilitating realist/anti-realist debate’ (p.81).
[5] See Hume (1777), Fricker (1995) on inferential experience and reductionism; Shapin (1994), Serjeantson (1999) on social and moral codes of conduct; Lipton (1991, 1998) on inference to best explanation.
[6] I do not deny the importance of an epistemic examination of the content of visual claims: within the scope of this paper a full discussion of imagistic content simply cannot be satisfactorily accomplished. For the purposes of this essay, then, concepts such as visual assertion, linguistics and misrepresentation will have to be assumed as part of the broad programmatic sketch suggested by an approach to illustrations as testimony.
[7] See Shapin & Schaffer (1985) on Boyle, literary virtual witnessing and seventeenth century efforts to efface the author’s role in verbal testimonial via experimental narrative.
[8] Coady (1992), 81.
[9] See Galison (1998) on particles; Dennis (1989) on microscopy; Alpers (1983), Harley (1988), Nuti (1999) on mapping; Galison (1998), Kemp (1996) on anatomical imaging; Lynch & Edgerton (1996), Schaffer (1998), Topper (1996) on astronomical imaging.
[10] The ‘instrument’ at question does not need to be extra-ocular: the informant’s own eyes may be instruments of scientific enquiry.
[11] See Dennis (1989) on Hooke and Winkler & Van Helden (1993) on Hevelius; Dalston & Galison (1992), Pang (1994, 1995) and Tucker (1997) on Victorian attitudes to photography; Galison (1998) on twentieth century ‘judgement’.
[12] Cultural concepts of seeing and knowing through images complicate this subject: see Alpers (1983).
[13] The examples do relay crucial information about their objects without providing ‘true-to-life’ pictures of them: see Bloom (1978) on Galileo and Harriot; Gombrich (1960), 74-83 on Dürer and ‘al vif’ drawing; Kemp (1990, 1997) on Stubbs.
[14] See Galison (1998) on solar fields; the British Library exhibition The Lie of the Land: the Secret Life of Maps (July 27, 2001 - April 7, 2002) on the elephant map.
[15] See Dennis (1989), Winkler & Van Helden (1993), Galison (1998), pp.345-347.
[16] It is precisely this glossy finish in scientific testimony that inspired a change in scientific reportage in the seventeenth century, along with the slogan nullus in verba. Yet artistic skill did not inspire such distrust in those who adopted the motto (i.e. early Royal Society): it was rather seen as a way through which a wider audience could share an empirical experience. See Dennis (1989), Shapin & Schaffer (1985).
[17] Lipton (1998). (See also Lipton, 1991).
[18] See Bender & Marrinan (2002), Hall (1996) on diagrams; Tufte (1983) on visual exposition of data.
[19] See Alpers (1983), Ashworth (1990), Daston & Galison (1992), Dear (1987) on the tensions between the emblematic or universal and the singular in art and science.
[20] See Watson, (1968).
[21] Exact replication of historical images has been seen as problematic for the image’s claim to truth-content. But within their historical and cultural context, exactly replicated images possess different and nuanced meaning, whether or not embedded in a stable external frame: see Gombrich (1960) esp. pp.58-73, Hall (1996), Ivins (1953), Kusukawa (1997, 2000).
[22] See Bloom (1978), Whitaker (1999).